W hat's behind the declines in violent crime? The question prompts lively discussion among people coming at a huge social issue from different angles: Some point to random demographic changes, others cite lock-'em-up prison policies; still others, most recently, point to more astute policing. This debate is not exactly a replay of the old argument over root causes versus tough law enforcement. The deep social pathologies that breed crime are still there, and that argument unfortunately remains on hold. Instead, the recent drop in crime rates poses a central strategic issue of criminal justice: Should it be reactive, emphasizing the capture, adjudication, and punishment of criminals after they commit crimes? Or proactive, working to prevent crimes from ever occurring? In principle, this should not be an either/or matter, but limited resources force choices.
Call it the "back-end/front-end" debate. Back-enders, focusing on events at the conclusion of the criminal justice process, favor punishment for its own sake and for its deterrent effects. They like the death penalty, long prison terms, and limited discretion for judges and parole boards who might be tempted to reduce them. Front-enders look for results from the early stages of justice: policing, gun control, drug treatment, and other kinds of alternative (to prison) programs for young offenders. In general, the back-end approach attracts conservatives who like to sound tough; the front-end approach attracts liberals who focus on broader social dynamics.
Which is the better way to fight crime? While the question ought to be pursued seriously-it is richly complex in practical, economic, and moral issues-it became hopelessly politicized during the decade that began around 1985, when crack and guns produced a surge of urban crime and politicians sought ways to exploit the fear it generated. As it turned out, this politics of crime heavily favored back-enders as it produced fervent support for capital punishment and a nationwide movement toward three-strikes and other mandatory-sentencing laws.
The movement was grounded in valid public anxiety about the level of crime, which no longer could be dismissed as an urban ghetto problem. Millions of middle-class Americans were waking up to the fact that fear had transformed their daily lives. Rising crime imposed surcharges for locks, alarms, and insurance; limited their use of parks, subways, neighborhood streets, and other public places after dark; forced complicated logistics for the supervision and protection of children.
Inevitably, people with narrow agendas sought to exploit the fear. Front-enders lamented the new "iron triangle" that lobbied relentlessly, and effectively, for harsher criminal sentences. Its three components: right-wing Republicans seeking to contrast themselves with "soft-on-crime" Democrats; builders, suppliers, and labor unions that benefited from expanding prison construction; and the National Rifle Association, which reflexively sought to fend off serious gun control with proposals for mandatory terms and sentence enhancements for crimes committed with firearms.
Front-enders sputtered in frustration as lawmakers brushed aside evidence that the fear-driven back-end agenda held no promise of greater crime control, and that it created something close to official racism as it forced disproportionate numbers of young black men into prison. What the front-enders failed to grasp was that the discussion had moved away from crime control, where it belongs, and into uglier, more primitive territory. Simply put, the frightened public gave up on government's ability to prevent crime and turned to other ways of handling its fear.
One of these ways was ad hoc privatization: small armies of security guards for hire; profitable new industries (the Club and Lojack to protect cars; cellular phones with buttons programmed for 911; more sophisticated alarm systems). Another fear-driven remedy was the demand for revenge, or, more precisely, for "expressive punishments" that put more emphasis on venting collective rage than controlling crime. Thus did huge majorities support the death penalty and longer prison sentences; in addition, millions applauded the caning of a young American for vandalism in Singapore and called for legislation to make corporal punishment possible here. They cheered as state lawmakers revived chain gangs and convict stripes and sought to eliminate the "amenities" of prison life. Sensing the public mood, lower-court judges toyed with public "shaming" as an alternative to jail for misdemeanants. Legislators, relieved that they could satisfy voters without having to control crime, were glad to go along with this use of criminal justice for mass therapy.
COST-EFFECTIVE CRIME PREVENTION
While a back-end strategy could guarantee a quick political payoff, no serious policymaker could ignore the longer-term costs. Between 1984 and 1994, according to the Federal Bureau of Justice Statistics, the number of convicts admitted to the nation's state and federal prisons in a year swelled 120 percent, from 246,260 to 541,434, boosting the total incarcerated 116 percent, from 419,346 to 904,647. The taxpayers' overall bill for criminal justice-police, courts, and corrections-also nearly doubled in the period, from $45.6 billion in 1985 to $93.8 billion seven years later, with corrections' share of the total increasing from 28.6 percent to 33.6 percent, or $31.5 billion.
What, in fact, was all this money buying? On this point, the statistics were hardly reassuring. The issue is one of scale. Perhaps half of serious crimes are reported to police. Of these, only about one-fifth result in an arrest. Less than two-thirds of those result in a conviction, and a tiny percentage wind up serving time in a state or federal penitentiary. Thus, the 20 million serious crimes committed each year produce about 500,000 incarcerations-and a third of them are for nonviolent drug offenses or drunk driving. Even if each convicted felon is responsible for many more than one crime apiece, how can incarceration of so small a fraction of serious criminals have much effect on the crime rate, either directly or as a deterrent? And, if budgets are limited, how is it possible to justify spending 33.6 percent of all the available money to impose serious punishment for a tiny percentage of serious crimes?
People determined to promote the back-end strategy point to studies that document crimes and costs apparently saved by incarceration. William Bennett and his co-authors John DiIulio and John Walters refer in their book Body Count to surveys of prison inmates in Wisconsin and New Jersey who claim to have committed numerous crimes in the year before their imprisonment. Both groups of inmates self-reported medians of 12 property or violent crimes, excluding drug crimes. The authors quote other research finding as many as 21 averted crimes for each incarcerated prisoner.
They also quote a study that sought to assess not only direct costs to victims but "monetary value of lost quality of life" caused by crime. "Using various measures," the study put prices on individual murders ($2.4 million each), rapes ($60,000), arson ("almost $50,000"), assault ($22,000), and robbery ($25,000). Multiplying numbers like that by the annual "crimes averted" factors found in the studies of inmates yields amounts that dwarf the average annual cost of keeping an inmate in prison (about $20,000).
However such calculations might provide ammunition for lobbyists of the iron triangle, they remain less than persuasive. Obviously, incarceration incapacitates criminals who are subject to it, and many criminals do commit many crimes per year. But the back-enders leave their audiences with an incomplete picture, for nearly everyone who goes to prison eventually gets released. And given the lack of rehabilitation resoundingly documented by recidivism studies over the years, most of those coming out can be expected to commit new crimes at similar rates. Thus, while 541,434 criminals were sent to prisons in 1994, 456,942 came out, for a net reduction that year of only 84,492 criminals. This does represent an increase over 1984, when 246,260 went in and 221,768 came out, for a net reduction of 24,492. But it's hard to see how incapacitating 60,000 more criminals, a figure that includes nonviolent drug offenders, can have more than a modest impact on serious crime rates even if one believes that each person incapacitated would have committed 10 or 20 crimes in a year. The net incapacitation figure, furthermore, is small enough to be overwhelmed by an increase in the number of young people recruited into lives of drugs, crime, and guns each year, as happened in the late 1980s. And, of course, as legislatures weary of spending tax dollars for prison expansion, allowing the surge of incarceration to level off, the figures will reverse for a time, with more people coming out than going in, for a net increase of criminals on the street.
As for the claim that the aversion of crimes saves society money, front-end strategies could save as much or more. In any case, estimations of crime control savings don't balance public budgets. And so far, the idea of saving taxpayers so much money in averted costs of crime hasn't led conservative back-enders to support hefty tax increases on them to finance more prison construction.
There simply is no escaping the troubling distortion of spending for corrections at the expense of the rest of the system. Couldn't some of the $31.5 billion that goes to lock up a few hundred thousand serious criminals for a few years each be put to better use preventing some of the 20 million serious crimes?
It's also instructive to think of the issue from the neighborhood's point of view. Suppose that crime may be reduced in equal measure and at equal economic cost either by putting a lot of people in prison or by putting more police on the street and developing other front-end programs to intervene with offenders early. Which strategy leaves the community better off?
Increased police presence risks increased abuse of civil liberties by overzealous officers, a problem that inflicts temporary aggravation on some innocent citizens. But sending people to prison inflicts severe, if not calamitous, emotional and financial stress on their innocent spouses, children, and parents. And their neighborhoods suffer the consequences of having to cope with ex-convicts as they return with their employment prospects permanently stunted and their ability to function in family and community life further impaired by the various brutalities of prison.
Other things being equal, the community clearly is better off with more police and front-end programs than with more people going to prison. In these terms, a strong case could be made for controlling crime with police even if it costs more than controlling it by sending people to prison. Any evidence that the police approach produces greater crime control at lower cost should blow the prison strategy out of the water.
OUNCES OF PREVENTION
In a sense, liberals who embrace a front-end law-enforcement strategy are growing up. A crime control agenda based on prevention arguably might include almost any measure that improves education, creates jobs, supplies day care, improves low-income housing, increases access to health care, and otherwise supports poor families. But by ignoring the citizenry's immediate anxiety about personal security, liberals who emphasized only "root causes" seemed hopelessly naive, and ceded the whole crime issue to conservatives. A front-end agenda of direct prevention doesn't mean giving up on root causes. But it does allow those with a more social conception of crime to embrace an approach that has two immense advantages over the back-end response: It is less vengeful-and more effective.
C onsider New York City: The violent summer of 1990 prompted the city's first black mayor, David Dinkins, and his police commissioner, Lee Brown, to push a proposal for new police hiring-and a tax to pay for it-through the city council and state legislature. The effective expansion of the department from 25,465 to more than 32,000 officers would turn out to be a gift of immeasurable value to Police Commissioner William Bratton, brought in from Boston by Mayor Rudolph Giuliani, who defeated Dinkins's try for a second term in 1993.
Bratton had previously served as chief of New York City's transit police, where he had experimented with new strategies. Now he returned with large ambitions that he would realize all too well, attracting so much attention for genuine achievement that the mayor, feeling upstaged, forced him out after two-and-a-half years.
Bratton's approach was to disperse responsibility for crime fighting downward to precinct commanders while instituting weekly meetings to hold them strictly accountable for results. In order to measure them, he forced precincts to produce a wealth of statistical data. Police computers began to map out crime and enforcement patterns with unprecedented precision and timeliness-they might show, for example, that reports of shootings on a certain street corner occurred mostly on Fridays and Saturdays after 9 p.m.
Precinct commanders were called to account in weekly "COMSTAT" meetings: Why are there so many shootings on that street corner? What do we know about that location and the people who frequent it? What are you doing to get it under control?
At the same time, the new commissioner found excellent use for the new cops coming out of the academy as a result of the Dinkins/Brown hiring plan. He ordered a citywide campaign against "quality of life" offenses-drinking in public, urinating on the street, making noise, and other forms of rowdiness.
Though the endeavor sounded like a public relations stunt, it was deadly serious, with purposes that ran far deeper than simply promoting better manners in public. Bratton was aware of the 1982 Atlantic Monthly article "Broken Windows," much discussed in police circles, by James Q. Wilson and George Kelling; it compares a neighborhood where police ignore low-level offenses with a building where the landlord ignores a broken window. As people realize that that they can get away with it, they begin to break more windows until the building is destroyed. Wilson and Kelling used the analogy to argue that determined policing of low-level offenses could inhibit serious criminal activity as well.
In New York, in 1994, "quality of life" became the excuse for an aggressive form of patrolling targeted on youthful lowlifes. It generated complaints of harassment even as it drew praise from older residents of troubled neighborhoods. The routine, based on police lawyers' careful study of Supreme Court "stop and frisk" decisions, called for officers to stop and request identification of anyone they suspected of committing an infraction, accepting only government-issued picture ID.
Those not carrying proper ID or found to be the subject of outstanding warrants were taken into custody, driven to the precinct station, and turned over to detectives who interrogated them for whatever they might tell about drug and gun trafficking and recent crimes in the neighborhood. The process added mightily to the flow of fresh information on which to base new operations.
The effects were immediate and dramatic. The number of homicides in the city had begun a gradual decline in the last years of the Dinkins administration. With the arrival of Bratton, COMSTAT, and aggressive patrolling, the homicide rate began a steep decline that appears to be continuing. Only 985 homicides occurred in the city in 1996, a decline of 57 percent from the peak number of 2,262 in 1990.
Bratton declared that he had proved the broken windows theory. His new measures, he said, had inhibited street criminals, causing them to leave their guns and drugs at home. These claims met with skepticism at first. Weren't crime rates going down all over the country? What was so special about New York? And weren't a lot of things beginning to happen that could be reducing crime independently of the police?
So far, the New York story survives those questions. The nation's big-city homicide rate turned down after 1991 and has continued to fall through 1996. But New York's decline exceeds the national figure. The homicide rate for cities of more than 1 million fell from 33 per 100,000 in 1991 to 21 per 100,000 in 1995. In New York, the rate fell from 29 per 100,000 to 16 per 100,000 in the same period.
Other explanations for declining crime include the natural maturing and waning of the crack epidemic, shifts in drug market patterns, and demographic changes that leave fewer crime-prone teenagers on city streets. But these more gradual events don't explain the close congruence of sharp declines in New York City's crime and the introduction of Bratton's new management and strategies.
Meanwhile, the only independent analysis so far, conducted by Andrew Karmen at John Jay College of Criminal Justice, offers some striking findings. Karmen found that homicides committed with guns and those committed out of doors fell more sharply than those committed indoors or with other weapons. He also found that homicides declined with the rise of patrol strength and with increases in misdemeanor arrests for quality-of-life offenses. Such findings bolster police claims that their greater numbers and aggressive patrols are inhibiting gun use and street crime.
The calculation of costs and benefits extends well beyond the criminal justice system and crime victims. A city experiencing dramatic declines in crime from policing, as opposed to slight or negligible ones from increased incarceration, becomes more hospitable to tourists and to businesses. And the good news, palpable on every street corner, calms the middle-class homeowners whose periodic bouts of panic about the city's future weaken their stabilizing commitments to neighborhoods and schools.
STREET-SMART INTERVENTION
While the New York experience is especially striking given the size of the city and its police department, it isn't unique. Bostonians have recently seen a drastic decline in crime, particularly in gun violence among juveniles. Observers credit a comprehensive police strategy characterized by unprecedented involvement with communities and cooperation among law enforcement agencies. Houston, Dallas, and San Diego have also seen big declines in crime, apparently the result of increased police presence and more aggressive patrolling.
The successes of police-based approaches to crime control encourage thinking about other front-end measures. At least three spring to mind immediately:
Invest more in lower courts, probation departments, and early alternatives. Offenders sentenced to state penitentiaries for serious crimes typically wind up there only after committing a number of lower-level crimes (only some of which come to the attention of the authorities) for which they receive insignificant sentences to lightly supervised probation or "time served"-days already spent in jail awaiting court action. Especially if they are relatively young, offenders get off with such sentences because judges are reluctant to expose them to the routine terrors of penitentiary life for a first or second offense.
There is broad agreement that much crime might be averted if courts were able to intervene with offenders more meaningfully after the first or second minor offense, rather than waiting for them to commit more serious crimes. Yet the nation's overcrowded, underfunded urban arraignment courts are a classic horror show of criminal justice. The heavy workload, burnout, and cynicism among criminal justice workers usually preclude any careful consideration of the offender and the underlying problems-substance abuse, lack of education, family crises-that lead people into low-level criminality.
During the 1990s, a few jurisdictions found the will and the resources to improve lower courts. Some set up "drug courts" where judges sentence drug-abusing offenders to treatment programs, then monitor their progress, retaining the power to incarcerate them for failure. New York City set up a somewhat different model in midtown Manhattan. This "community court" arraigned low-level offenders of all sorts, sentencing them to community service projects in the neighborhood, and referring them to a well-staffed social service office located on the premises.
Innovative efforts of lower courts are enhanced when probation departments are able to help. Agencies that deal with offenders released under court supervision are likely to be as under-resourced as lower courts. Exceptions are found in Phoenix, Arizona, and in the state of Georgia. Both places offer judges probation-managed "ladders" of sanctions-sentencing options that increase in severity from standard probation supervision through "intensive supervision" (lower caseloads), electronic monitoring, and house arrest, up to work release and boot camp programs based in secure residences. Judges greatly appreciate the chance to move offenders up and down the ladder as they demonstrate more or less willingness to behave.
The possibilities are enhanced further as probation departments get creative with alternative sanctions, finding politically acceptable modes. In South Carolina, for example, judges sentence offenders to pay victims restitution rather than serving time in prison. Offenders who don't have a way to pay are sent to secure residences on the grounds of state prisons, then bused out to work each day in private-sector jobs until they earn enough to pay off their sentence amount. Job developers at the centers come up with the placements, typically hard-to-fill minimum-wage slots where employers frustrated by high turnover welcome the restitution center's steady supply of workers, who are dependably sober, drug free, and motivated by the desire for release.
The restitution centers suggest the potential for public acceptance of front-end programs. Reliable payments of victim compensation defuse much criticism of the non-prison sanction, while the restitution workers' value to local employers builds support for the idea in the business community.
Expand drug treatment. The link between substance abuse and criminal behavior remains obvious and research suggests that as drug abusers recover from addiction, they recover from criminality as well. In large measure, the success of a front-end strategy that calls upon judges and probation agencies to do more with offenders in the early stages of their criminal careers depends upon abundant availability of drug treatment.
The goal should be to develop enough treatment slots so that all addicts who voluntarily seek help may obtain it immediately, and so that judges who wish to make treatment part of a sentence package can order an offender to begin at once. This could be accomplished without any need for big new federal or state bureaucracies simply by amending the Medicaid law so that it will reimburse drug addiction therapy provided through free-standing programs rather than in hospitals.
Skeptics point out that the treatment programs have low rates of success. How can one be sure money spent on them doesn't go down the drain? Yet programs that move even, say, 25 percent of clients into long-term recovery may wind up costing less than sending the same offenders to prison for short terms, then returning them to lives of addiction and crime. Furthermore, treatment managers say that an addict may need several attempts at treatment before it "takes." As courts require offenders to try again and again, the success rate increases.
Get serious about gun control. Researchers confirm the police belief that guns in the hands of kids played a central role in the burst of crime that began in 1985. Even so, lobbying of the National Rifle Association minimized new gun-control legislation during the 1980s and early 1990s. The Brady Bill and the ban on assault weapons, hailed as big symbolic victories, are relatively modest measures. In 1994 and 1995, however, the NRA overreached politically and, by some accounts, financially as well, and its influence began to recede. That makes real gun control look feasible.
A serious gun policy would, at a minimum, require as much of a person who wishes to own and use a gun as of one who wants to own and use a car. Guns should be numbered and registered, with data on guns and owners stored in a computer database instantly accessible by law enforcement agencies.
In addition, gun owners should be licensed, and the burden should be on applicants to demonstrate that they are mentally healthy, have no criminal or spousal abuse records, and have no problems with drugs or alcohol. They should be required to pass written tests on gun law and gun safety as well as practical tests of gun handling on a firing range. And they should have to carry substantial liability insurance. Beyond that, Washington could require manufacturers to build in safety devices like trigger locks that permit use of the gun only by the registered owner, and magnetic strips or computer chips that make guns easier to detect and trace.
Finally, federal law could require that anyone who wants to purchase more than one gun per month make the case for such a need to the local police. Such a law poses no inconvenience to virtually all legitimate gun purchasers, but it could severely inhibit profiteering by gun runners who make legal purchases from retail stores and resell the weapons illegally on the street.
H ow much would a front-end strategy cost? Obviously the expansion of police departments, lower courts, probation agencies, alternative sanctions, drug treatment, and the bureaucracies necessary to enforce new gun laws would require significant spending. The back-end strategy, however, has already committed the nation to billions in new spending as prisons expand and courts fill them with tens of thousands of new inmates. The issue may not be one of coming up with new money so much as engineering a partial shift of funds already in place.
For now, it's enough that police-led victories over crime in New York and other cities revive the front-end/back-end debate and demonstrate an urgent need for research: How are more police used most efficiently? How can lower courts adapt themselves for early intervention and crime control? What are the optimum staffing levels for probation departments? What kinds of alternative sanctions yield the best results? How do different modes of drug treatment work for different kinds of addicts? What would national gun registration and licensing entail and what would be their likely effects?
Such questions, considered marginal where back-end assumptions dominate talk of criminal justice, now belong at the head of the agenda.